How should Cisco CallManager virtual appliances be represented in CMDB?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
3 weeks ago
Hi everyone,
I am working on Cisco CallManager Discovery and have reached a point where I could use some guidance.
The Cisco CallManager pattern connects through SSH and successfully retrieves information such as the hostname, version, node role, and cluster members. The failure happens when ServiceNow tries to create the final application CI because it cannot find a suitable underlying host CI and relationship.
Normal UNIX Discovery does not work for these appliances because the SSH account opens a restricted Cisco command interface rather than a standard Linux shell.
My current understanding is that I may need to discover the underlying virtual appliance first and then relate the Cisco CallManager application CI to it.
For anyone who has implemented this successfully, which CMDB class did you use for the underlying appliance, and which relationship did you use between the appliance and the Cisco CallManager application?
I have already confirmed the SSH connectivity and pattern output. I am mainly trying to understand the correct CMDB model before making any customization.
Thank you.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2 weeks ago
Hi AbdulBadi,
Your understanding is correct - the model is the standard "application on a host" two-CI model, and the missing host is precisely why the pattern cannot finish:
1. Application CI - the Cisco CallManager pattern creates the application CI in the Cisco CallManager class (cmdb_ci_appl_cisco_callmgr, under Application). One application CI per CUCM node - publisher and each subscriber get their own.
2. Host CI for the appliance - use Computer (cmdb_ci_computer), not Linux Server. The Linux Server class implies an OS that Discovery can interrogate; with the restricted ADE-OS shell it cannot, and forcing the Unix classifier is what produces the broken/partial hosts. The host CI represents the appliance itself: name = the CUCM node hostname, IP address populated, Virtual = true. This is the CI the pattern's host-dependency check looks for - it resolves the host by the IP/hostname it connected to, so the name/IP on the Computer CI must match what the pattern reaches.
3. Relationship between the two - Runs on: Cisco CallManager application CI --Runs on::Runs--> Computer host CI. When the host CI exists beforehand, the pattern creates/uses this relationship itself; you do not need to build it manually.
4. Relationship to VMware - the Computer CI is the guest, so link it to the VMware Virtual Machine Instance (cmdb_ci_vmware_instance) with Virtualized by (Computer --Virtualized by--> VM Instance). The clean way to get both the host CI and this relationship right is sequencing: run vCenter discovery first. vCenter discovery creates the VM instance and the virtual Computer CI for the guest (identified by the VM's UUID/serial), and once that Computer CI exists, the CallManager pattern stops failing on host dependency and attaches the application CI to it on the next IP-based run. If a node is not in vCenter discovery scope, create the Computer CI manually (or by import) with the correct hostname/IP and virtual flag, relate it to the VM instance, and the pattern will adopt it the same way - the one thing to get right is the identifier: give the manual host the VM's serial/UUID where you can, so a later vCenter discovery run reconciles to the same CI instead of creating a duplicate computer.
What not to do: do not model the appliance only as a VM instance (the application pattern cannot hang off cmdb_ci_vmware_instance - it needs a Computer-class host), and do not create a custom "appliance" class for the host; cmdb_ci_computer with Virtual = true is the supported representation and keeps identification, relationships and the pattern's expectations standard.
Regards,
Karimuddin
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2 weeks ago
- last edited
2 weeks ago
by
kh-christelle
Hi @kmohammed
Thanks for replying, I am still bit confused on this. Would you mind sharing a way I can reach out to you to get more clarity on this issue that I am having?
Thanks,
Abdul Mohammed
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2 weeks ago
Hi Abdul — I don't take discussions off-thread (no direct messages), but let's clear it up right here so others with the same issue can follow it too.
To help pinpoint the confusion, could you tell me which part is unclear?
1. The class choice — why the appliance host is Computer (cmdb_ci_computer) rather than Linux Server, even though the appliance runs a Linux-based OS?
2. The relationships — how the CallManager application CI (Runs on) the Computer CI, and how that Computer CI then relates to the VMware VM instance?
3. Or the discovery order — what the pattern actually created in your instance so far, and where it stops or errors?
A quick recap in one line: the pattern creates the CallManager application CI; the appliance it runs on is modelled as a Computer CI (the guest), and that Computer CI is linked to the VMware VM instance that hosts it. It is not modelled only as a VM instance, because the application pattern needs a host CI to attach to.
If you share what your discovery log / created CIs show now (and which step above does not match what you see), I can walk through that exact step here.
Regards,
Karimuddin